are URIs identifiers?

[Joaquin Miller]

Gabe wrote:
>
>This is good stuff, Joaquin, but it seems to be at odds with the use 
>of term "identifier" in the URI/web context.
>
>Any string that conforms to the URI syntax (RFC 3986) is defined as 
>a URI (a type of identifier). There's no need for that URI to refer 
>to exactly one entity. The URI exists simply by minting it (writing 
>it down). Thus, I don't see how that aligns with the statement 
>(summarizing your summary) that an identifier is a name in a context 
>that "refers" (what does *THAT* mean) to a single entity.

Thanks, Gabe.  I don't deny that it seems to be at odds.  If it    i 
s    at odds, then i am wrong.

I won't rise to the refers bait (though it sure is well placed!), but 
if we can take    that   as understood, here is how i see the situation:

The namespace of URIs is defined by the URI syntax.  Any string that 
fits the syntax is a name in the URI namespace.  [RFC 3986 : 3]

"A Uniform Resource Identifier (URI) is a compact sequence of 
characters that identifies an abstract or physical resource." [RFC 
3986 : Abstract]

I feel that sentence says a sequence of characters is not an URI 
unless it identifies a resource.

But let's read on, and give a sympathetic reading.

I feel that Clause 3, when it says things like "Each URI begins with 
a scheme name that refers ... ," means exactly that: each URI 
conforms to the syntax.  It does not mean that each sequence of 
characters that conforms to the syntax is a URI.  When Clause 1.1 
says  "A URI is an identifier consisting of a sequence of characters 
matching the syntax rule named <URI> in Section 3," it means just 
that: every URI matches the rule, not every sequence of characters 
matching the rule is a URI.

But I feel it is the intention of the RFC that that each sequence of 
characters that conforms to the syntax is a name in the URI 
namespace.  It's simply that every name in that namespace is not 
necessarily a URI.  I feel that fits with the X.900 definition: 
"2-12.3 Name space: A set of terms    usable    as names."

Back at Clause 1.1, we find: "An identifier embodies the information 
required to distinguish what is being identified from all other 
things within its scope of identification.  Our use of the terms 
'identify' and 'identifying' refer to this purpose of distinguishing 
one resource from all other resources, regardless of how that purpose 
is accomplished (e.g., by name, address, or context).  ... Likewise, 
the "one" resource identified might not be singular in nature (e.g., 
a resource might be a named set or a mapping that varies over time)."

I feel that the X.900 definition works with this.  That a name refers 
to one entity at a certain time and at another at a different time 
works with "2-12.2 Identifier: An unambiguous name, in a given naming 
context."  Such a name is not necessarily ambiguous.  Example: 
'President of the United States'.  Our system of laws, here in the 
middle of North America, is based on that being an unambiguous name 
in the context of our system of laws.  We have a special ceremony to 
ensure there is no ambiguity.

And certainly a collection of entities may be an entity.  Thus 
'Senate of the United States' is, in the context of our Constitution, 
an unambiguous identifier for a collection of one hundred persons 
(sometimes less, never more).

Obviously folks sometimes say 'URI' when they mean "sequence of 
characters matching the syntax rule."  And that's fine in a 
conversation where precision of terminology is not required or even 
useful.  And it is fine in a conversation where it gives as much 
precision as is needed.  But it helps to have a well defined jargon, 
and to stick to it when it is useful.

It's fine for me and you (no insult intended, if you are a sailor) to 
refer to a line using 'rope', but if we are on deck in a storm, it is 
not best for us to holler to the captain through the wind "A rope has 
broken", when she needs to know immediately whether the line that has 
parted is a sheet, a halyard, a shroud, something else essential, or 
merely a ratline.

Cordially, Joaquin

The answer to the question in the Subject is: Yes.


>... there are no identifiers except in relation to a naming 
>context.  With respect to a given naming context, an identifier is a 
>name for a single entity.  If, in a given naming context [and at the 
>same time], a name refers to two or more entities, that name is not 
>an identifier.
>
>Cordially, Joaquin
>
>I feel it would be a waste of our time, but if challenged I can 
>quickly demonstrate that there is no such thing as an identifier 
>that is unique.  My point about 'unique identifier' is that it 
>muddles the conversation and sows confusion.  But another point is 
>this: It's not useful to call an identifier that is not unique a 
>'unique identifier'.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-general/attachments/20060920/2c6accc7/attachment-0001.htm>